ECE50024_FinalProject_Asha

Project on Graduated Non-Convexity for Robust Spatial Perception:

Quick Start: Open Python- directly run gnc_gm_pcl.py code file. Change the outlier percentage and the output will print transformations.

For GNC_RANSAC_forPCL.py, first run the 1st section i.e GNC for Point cloud registration and then run 2nd section using RANSAC Method. Lastly run code for plots which will print diiferent transformation error plots w.r.t outlier percentages.

For GNC_LP_ForLinearRegression.py file similarly run the 1st section to define problem and then Run the other sections in order which will print the fit curves as output. As outlier percentage and number of datapoints are taken as input so by changing them the performance of the methods can be observed.

Run Robust_cost.py code to plot the robust cost function w.r.t to residual and changes in the shape of curve with change in mu.

The plots and graphs are also provided in this repository in .png format.